Bump edition 2021 -> 2024. rust-version stays at 1.91, already well above
the 1.85 floor the edition needs; docs/guide/resources/troubleshooting.md
still claimed 1.70+, which is where a failed `cargo install --git` lands.
Four things the edition forces:
- std::env::{set_var,remove_var} are unsafe in 2024 with no safe std
replacement. Rather than wrap the test call sites in unsafe -- which the
crate denies and .semgrep.yml flags -- route them through temp-env, a
dev-only dependency whose closure API is safe and which restores the
previous value even when the body panics. The hand-rolled CLAUDE_DIR_LOCK
and PI_DIR_LOCK guards existed only to serialise those mutations and are
now redundant; CWD_LOCK and TEST_ENV_LOCK stay, they order more than the
env var itself.
- unsafe_op_in_unsafe_fn is on by default, so the libc calls in the proxy
signal handler and in stream.rs's relay handler need explicit unsafe
blocks, scoped to the libc calls themselves.
- `gen` is a reserved keyword, so the closure by that name in diff_cmd.rs
becomes make_lines.
- Tightened tail-expression temporary scopes let clippy prove the binding in
setup_test_env is inlinable, so let_and_return now fires there.
if_let_rescope changes when the scrutinee temporary drops in an if let/else.
The two sites in show_claude_config take cargo fix --edition's match rewrite,
which keeps the 2021 drop timing.
rustfmt.toml is kept rather than dropped: cargo fmt passes --edition from
Cargo.toml, but a bare rustfmt invocation has no crate context and falls
back to edition 2015, which cannot parse the let-chains the next commit
introduces. Pinning it there keeps format-on-save and pre-commit hooks in
agreement with CI.
clippy::collapsible_if is allowed crate-wide for now; the follow-up commit
adopts let-chains and removes the allow.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
`get_by_command` moved to the volume-weighted rate while this branch was
open (#891), summing every call and guarding the division at group level.
The two telemetry queries now use the same shape: the row-level
`WHERE input_tokens > 0` is gone, so a call with no input that still
printed something counts against its command exactly as it does in the
By Command table, and `HAVING SUM(input_tokens) > 0` keeps commands that
never had any input out of both, as before.
`low_savings_commands` lists net-regressing commands: `sav > 0.0` became
`sav <> 0.0`, so only exact 0% stays with `passthrough_top`. A command
that emits more than it saves is the filter to fix first, and with the
weighted rate the old bound also dropped commands the unweighted query
used to list.
The two regression tests opened the on-disk tracker, so every `cargo test`
wrote ten synthetic rows (one at 100k input tokens) into whichever history
database `Tracker::new()` resolves to, on a developer machine the real one.
They now use `Tracker::new_in_memory()` like the rest of the module, and
`test_avg_savings_per_command_inner_weighted` only asserted a `0..=100`
range that the unweighted query also satisfied. Both tests now assert the
exact weighted figures, check each listed rate against `get_summary()`,
and fail when either query goes back to `AVG`, to the row filter, or to
the positive bound.
Tests are numbered 18 and 19 after the ones develop already carries as
14-17. The `low_savings_commands` doc no longer points at a PR, and both
telemetry field tables describe the new semantics.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
The section landed in docs/TROUBLESHOOTING.md, a flat file removed in
a94e9493 when the docs were consolidated. Content there does not reach
the published guide, and the copy reintroduced install guidance that
develop had already pinned to --branch master.
Move the answer into docs/guide/resources/troubleshooting.md and drop the
duplicated Type Kit collision section, which that guide already covers.
The cache write/read breakdown is reported by `rtk cc-economics`, not by
`rtk gain`, so point readers at the command that actually shows it.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Restore to develop the files where the rescoping added noise without
helping a reader: all of .claude/, the src/ module READMEs that only
described their own filters, ARCHITECTURE.md, TELEMETRY.md, quick-start,
configuration and troubleshooting.
Also drop the "Two estimators, one caveat" table from savings-explained.md.
The page already states that rtk gain estimates bytes/4 and ships no
tokenizer; enumerating the test-side estimator was detail no reader needs.
The user-facing surface keeps the rescoping: READMEs in seven languages,
the guide, hooks, and the analytics and usage pages.
The arrow diagram read as a pipeline where bash output becomes input tokens
which become cost. The real relationship is containment: bash output is
part of input tokens, and input tokens are part of cost alongside output
tokens.
Replace the arrow chain with a tree in all 12 places it appeared, including
the six translated READMEs:
Cost
├─ Input tokens
│ ├─ Bash output <- the only part RTK filters
│ ├─ Your prompt
│ ├─ System prompt
│ └─ Conversation history
└─ Output tokens <- what the model writes
This also makes the dilution self-evident: RTK shrinks one leaf, so the
effect on the root is bounded by that leaf's share.
The documented release blocker was ">=60% savings", but the code never
enforced that. 23 of the 63 threshold assertions in src/ already sit below
60% (15, 20, 30, 40 and 50% appear across aws, gh, git, glab and mvn), so
the gate was aspirational rather than real.
Set the floor to 20%, pointing at the existing "Correctness VS Token
Savings" section for the reasoning rather than restating it: a modest, safe
reduction beats an aggressive one that drops information the agent needed.
Descriptive "60-90%" ranges are left alone where they report what filters
actually achieve. Only requirement statements moved to the 20% floor.
Also reduce links to savings-explained.md down to one per file, and only
from root docs and user-facing pages. Removed from docs/contributing/ and
.claude/rules/, where the surrounding text already carries the caveat.
RTK was documented as delivering "60-90% token savings", which reads as a
cost reduction. What RTK actually reduces is bash output bytes. Those are
one contributor to input tokens, which are themselves only part of a bill
that also counts output tokens, so the reduction dilutes at every step.
- add docs/guide/resources/savings-explained.md as the canonical explainer:
the savings chain, both estimators, and what RTK does not reduce
- rescope the headline claim across README (7 languages), the guide, hook
rules, agent definitions and module READMEs
- relabel per-command tables as bash output reduction, keeping every figure
- document that reported tokens are estimates: rtk gain uses bytes/4
(src/core/tracking.rs), filter tests use split_whitespace().count().
Neither is a real tokenizer, so ratios hold but absolute counts do not
Remove figures that had no source: the $3/Mtok constant and its $36
example, the +/-10% tokenization accuracy claim, the 99.5% hook-install
figure, the invented session tables in README and INSTALL, and the 30-50%
parser range.
CHANGELOG is untouched. Shipped release notes stay as a historical record.
Nicolas Le Cam